It has been a long time since I have spoken at the business analytics conference. I am now retired and working with a startup, Optimal Dynamics, that is taking my decades of research into models for the truckload trucking industry to the field. This work has required completely rethinking how to model and solve what I call "sequential decision problems" (decision, information, decision, information, ...). Sequential decision problems arise in every aspect of business, from managing physical resources (inventories, people, equipment, facilities), financial resources (cash, assets, insurance contracts, bank loans), and information resources (running field experiments, testing new ideas, marketing and advertising).
The academic communities that work on sequential decision problems tend to focus on the most complicated methods for making decisions (e.g. using Bellman's equation or stochastic lookahead policies), which are rarely (if ever) used in practice. I have been able to organize every possible method for making decisions, including anything proposed in the literature or used in practice, into four classes of policies.
At Optimal Dynamics, we use all four classes of policies to optimize truckload fleets (see attached video). This is not just academic games - decades of working with real companies through CASTLE Lab at Princeton University taught me the importance of realistic models, include the importance of modeling uncertainty. I was a Franz Edelman finalist twice, and won the Daniel Wagner prize for our work in truckload trucking that was then licensed to Optimal Dynamics.
I came out with a major new book last year (see the webpage at https://tinyurl.com/RLandSO/), but this is for people who actually want to write code (I have given tutorials on this to analytics groups at several major companies). I recommend starting with a book that I wrote for my undergraduate course that uses a teach-by-example style. The book can be downloaded from https://tinyurl.com/PowellSDAMbook/. For a quick introduction, see the video at https://tinyurl.com/sdafieldyoutube/
I am working on a new book that applies this framework to supply chain problems - the tentative title is: Supply Chain Analytics: Realistic Models for Real-World Problems. I would be happy to present this work at next year's analytics conference!
------------------------------
Warren Powell
Chief Innovation Officer, Optimal Dynamics
Professor Emeritus, Princeton University
http://www.castlelab.princeton.edu/------------------------------
Original Message:
Sent: 05-01-2023 11:10
From: Zohar Strinka
Subject: Favorite moments from the Business Analytics Conference
Hello Everyone! I hope you all had a great time at the INFORMS Business Analytics Conference last month in Colorado and got to make some valuable connections. I would love to hear about your favorite parts of the conference. Was there a specific talk that motivated or inspired you? A new tool or technology you learned about? A session you will be sure to go back to next year? Share your thoughts here!
------------------------------
Zohar Strinka
Analytics Strategies
Denver, CO
------------------------------